Written by:潮向研究

On June 1, Goldman Sachs released its daily Asia Pacific stock review "The 720", featuring a long list of names including Samsung, SK Hynix, Kioxia, MediaTek, Lenovo, and BYD. It looks like a comprehensive shopping list, but upon reading it, one finds an absolute core: storage chips.

The most significant judgment from Goldman Sachs in this issue is that the current storage upcycle "will last longer" (higher for longer), with shortages extending all the way to 2028, and the market has greatly underestimated its duration. The evidence lies in the valuations: most storage stocks are still trading in the mid-single-digit P/E ratios, indicating the market's perception that this is just another ordinary cyclical rebound, which Goldman does not agree with.

To break it down by importance, here is an overview with a quick reference chart at the end.

Main event: Storage shortages to persist until 2028, three companies collectively upgraded

Goldman Sachs compared this cycle with past ones, concluding that this time is different. There are three reasons: higher visibility of AI server demand, limited supply growth, and increasingly binding long-term supply agreements (locking orders and prices). With these three factors combined, the supply and demand for DRAM, NAND, and HBM in 2027 will be tighter than in 2026, with shortages continuing into 2028.

The most direct evidence is Goldman Sachs' DRAM supply and demand chart. Negative numbers represent supply falling short of demand; the deeper the gap, the more price support there is. This time, Goldman has revised its predictions for the period from 2026 to 2028 down to a deeper shortage range, with 2027's forecast changing from -2.5% to -5.9%, almost doubling. To put it plainly: Goldman believes that memory chip manufacturers will face increasing shortages next year and the year after, which means that price increases can last longer.

Specifically, three companies have been collectively adjusted:

• Samsung Electronics: Target price raised to 480,000 KRW for 12 months, maintaining a buy recommendation.
• SK Hynix: Target price raised to 3,500,000 KRW for 12 months, maintaining a buy recommendation.
• Kioxia: Upgraded from hold to buy, new target price of 93,000 JPY.

Kioxia is the only one with a rating upgrade this time, and Goldman’s reasoning is worth noting separately: it believes that the profit peak of this cycle will be higher than previously expected and can sustain for two to three years, rather than having a rapid fall. Based on this, Goldman has raised its operating profit forecast for Kioxia for fiscal years 2027 to 2029 by 16% to 48%, and expects the gross margin to remain at a high level of around 80%. Providing a three-year sustainable high profit forecast for a company involved in such a cyclical business is a strong statement.

AI computing power ecosystem: From chips to optical modules to data centers

Aside from storage, this issue nearly covered the entire AI hardware supply chain in China and Asia, unified by a single logic: global cloud providers' (hyperscalers) capital expenditures are accelerating, and money is flowing downstream along this chain.

• MediaTek: Buy, new target price of NT$5,000. The highlight is its transformation from mobile chips to data centers and custom ASICs (AI chips tailored for specific clients). The company aims to achieve $2 billion in data center/AI ASIC revenue by 2026, and capture 10% to 15% of the $70 billion to $80 billion ASIC market in 2027.
• Eoptolink: Buy, target price raised to 841 CNY. It manufactures optical modules, which are key components responsible for high-speed data transmission in AI data centers. Goldman is optimistic about its 1.6T optical module ramping up production in Q2 and accelerating in the second half of the year, and it has raised its profit forecasts for 2027 and 2028 by 5% and 6%, respectively.
• Biren: Buy, target price raised to HK$70.7. A domestic AI chip manufacturer, their Bili166 has received a high security rating. Goldman expects it to transition to higher-performance AI chips, becoming more expensive; in 2027, it can turn a profit, and revenue forecasts for 2026 to 2030 have been raised by 4% to 28%.
• Huaqin Technology: Buy, as it debuts in this issue. Target price of 149 CNY for A shares, and it covers its H shares for the first time with a target price of HK$127.76. The logic is that it is transitioning from consumer electronics OEM to AI data centers, with a projected compound annual growth rate of 32% in revenue from 2025 to 2027.
• Data center giants: GDS maintains a buy but lowers the ADR target price to $49 (due to slow onboarding speed and declining monthly service revenue, partially offset by the higher valuation of overseas DayOne business); CenturyLink maintains a buy, and its target price is raised to $16 (first quarter results exceeded expectations, capacity ramp-up execution is strong, suppression factors from strategic investors have been alleviated).
• Lenovo: Buy, target price raised from HK$27 to HK$31. This is betting on an AI PC replacement wave; Goldman expects its laptop market share to expand to 28% by 2028, with AI laptop penetration reaching 66%, raising the overall average price. Its profit forecasts for Lenovo for the fiscal years 2027 and 2028 are 22% and 25% higher than Bloomberg consensus expectations, showing a significant divergence.

Products that are not in the AI main line but are also mentioned

• Chinese real estate (China Overseas, China Resources Land): Goldman is assessing whether the rebound in the real estate sector can hold. It hypothesizes an optimistic scenario where 15 key cities follow the price recovery of Shanghai and Shenzhen, with prices rising by 15% by the end of 2028. Based on this premise, it estimates that China Overseas (COLI) and China Resources Land (CR Land) could expand cash profits by more than 30% and 50% respectively by 2028. Based on segment valuation, Goldman gives further upside potential of 52% for COLI and 76% for CR Land, and maintains a positive outlook on these two stronger state-owned developers. It is essential to highlight that this assessment is based on an optimistic assumption, not a baseline forecast.
• BYD: Buy, target price of 137 CNY/134 HKD. The highlight is its presentation at the smart strategy conference where the "Tian Shen Zhi Yan B" urban navigation assistant (NOA) is packaged as an optional accessory for 12,000 yuan, pushing the entry-level model price to 78,800 yuan, making it the cheapest NOA vehicle in China. They also launched their first self-developed 4nm smart driving chip "Xuan Ji A3", which is already in mass production. Goldman believes these engineering capabilities will increase the penetration rate of high-end smart driving, lower costs, and improve profit margins.
• Japanese semiconductor equipment: Goldman maintains buy ratings for Lasertec, Ebara, Disco, and Tokyo Electron. The only contrary action is to downgrade vacuum equipment manufacturer Ulvac from buy to neutral, with a target price lowered to 9,400 JPY, citing weak high-margin power semiconductor orders and slower-than-expected gross margin expansion.
• Panasonic HD: Buy, target price raised from 4,000 JPY to 4,220 JPY, optimistic about generative AI-related businesses (backup power, copper-clad laminate CCL, high-performance capacitors).
• NTT: Buy, target price slightly raised from 176 JPY to 179 JPY, focusing on domestic IT service demand and providing a safety margin for about 5% total shareholder return.

A macro mainline: AI prosperity meets energy crisis

Connecting the individual stocks is Goldman’s macro judgment: Emerging markets are being torn in half by two forces. On one side is the AI investment boom, and on the other side is the energy supply contraction caused by the blockade of the Strait of Hormuz.

Tech-exporting economies like South Korea and Taiwan benefit from the surge in exports and current account surpluses; energy-importing countries face rising inflation, currency depreciation, and fiscal strain from fuel subsidies. Goldman expects the average price of Brent crude oil in the fourth quarter to be $90 per barrel, continuing to pressure economies heavily reliant on oil imports, and recommends overweighting stocks in China, Korea, Brazil, and South Africa. This line connects well with recent developments in Iran and the macro backdrop of oil prices.

Additionally, two points that have a direct impact on the funding situation in A-shares:

China's imports surged by 23.6% year-on-year in the first four months of this year, but Goldman believes this is a highly concentrated phenomenon, with gold and semiconductors accounting for about 65% of the import increase, not reflecting a continuous deterioration in external balance.

The semi-annual adjustment of the CSI and CNI indices is estimated by Goldman to bring more than $48 billion of two-way passive capital flows, with the most inflow into the technology hardware and semiconductor, and capital goods sectors (approximately $3.1 billion and $1.4 billion respectively), while medical and banking sectors see the most outflow. New stocks expected to receive the most passive capital inflows, as pointed out, include Huagong Technology, Yuanjie Technology, Huahong Semiconductor, Zhaoyi Innovation, and Chipone. For funds engaged in index adjustment arbitrage, this is a clear signal.

Lastly, Goldman has included a little fun fact: the World Cup championship probability forecast for 2026 has Spain at 26% in the lead, followed by France at 19%, Argentina at 14%, Brazil at 8%, and England at 5%. The model has deducted points for defending champion Argentina, just for some entertainment.

Quick reference chart

This article is a整理与解读 of third-party brokerage reports by Deep Tide TechFlow. The ratings, target prices, profit forecasts, and related judgments quoted in the text are solely the views of the brokerage analyst and represent the position of their respective institutions, not the views of Deep Tide TechFlow, nor do they constitute any investment advice.

When reading, please note the following three points:

1. The target price is the analyst's expectation for the future period (usually 12 months), is a forecast rather than a commitment, and will be repeatedly adjusted according to the company's performance and market environment.

2. Sell-side research reports tend to be bullish by nature. It is common for brokerages to give a "buy" rating to covered companies, and some covered companies have interest relationships with the brokerage through investment banking activities. A list skewed towards "buy" should be read with this background in mind.

3. The value of research reports lies in their main logical framework and the underlying assumptions on which they depend, rather than a single target price. When the main line holds, the logic for related targets stands; if the main line is invalidated, the entire list of targets may wobble. Focus on the logic, not just the price.

The market carries risks, and decisions must be made independently. This article should not serve as a basis for buying or selling any securities.